[evlatests] Modcomp-Free (at last!)

[Rick Perley]

    After some troubles whose resolutions are best described by others, 
we were finally able to get an extended test of modcomp-free 
observations.  I did my usual 'stress test', with continuum and line, at 
all bands (including U-band).

    I report here only on X-band, continuum.  Others will follow 
tomorrow, as time permits. 

    1)  Top-Level Report

    Excellent data, generally as good as that taken with 'Modcomps on'. 

    27 antennas gave stable fringes, in all four IFs.  (I think this is 
a first!)  In particular:f

    The missing IFs in antennas 21 have been found.  The 'sinusoidal' 
phase in that antenna in RCP was not visible. 

    There were no phase jumps in any antenna. 

    2)  Some more detail on problems noted.

    The most important discovery is the difference in the phase 
residuals, on the long spacings. between EVLA and VLA antennas.

    a)  All VLA antennas have a phase ramp on a 10-second period.  
Without calculating the derivatives myself, it looks to me like the 
first derivative of phase is not being applied.  The ramp amplitude is 
proportional to baseline length, and changes sign between the E and N 
arms.  It was barely visible on the W arm.  The amplitude is large:  40 
degrees at E64 (w.r.t. antenna 9, at the center of the array).
    b) All EVLA antennas have parabolic phases, on the same 10-second 
period -- indeed, these are the same as seen with Modcomps in control.  
The amplitudes are less than the VLA phase ramps -- a maximum of 30 
degrees, trough to peak, at E72.  The 'cups' also change sign between E 
and N arms, and they are not visible on the W arm antennas.  This 
problem looks like a misapplication of the 2nd derivative of phase. 

    Otherwise, the usual litany of small troubles were noted:

    Short shallow drops were seen on 19 and 26, in IF 1 only. 
    26B is unstable (probably related to Tsys fluctuations)
    Antenna 24B  has reverted to the 'multiple amplitude' state we've 
seen in the past! 
    Antenna 14, IFs C and D are unstable -- due to Tsys fluctuations. 

    A number of antennas give low amplitudes -- we are certain this is 
due to large pointing residuals, which are being measured, as I type this. 

    All in all -- a significant improvement over recent atttempts.